A new technique developed by researchers at Shanghai Jiao Tong University and other institutions enables large language model agents to learn new skills without the need for expensive fine-tuning....
A new technique developed by researchers at Shanghai Jiao Tong University and other institutions enables large language model agents to learn new skills without the need for expensive fine-tuning....
Bottom line: CORE was more efficient than the baselines on each dimension: fewer rollouts, fewer training samples, and fewer added context tokens.
Bottom line: CORE was more efficient than the baselines on each dimension: fewer rollouts, fewer training samples, and fewer added context tokens.
In the 10-example train setting, CORE beat every baseline’s best result (after 1000s of rollouts) within its first 350 rollouts.
In the 10-example train setting, CORE beat every baseline’s best result (after 1000s of rollouts) within its first 350 rollouts.
🔗 aidailypost.com/news/memrl-b...
🔗 aidailypost.com/news/memrl-b...
MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic Memory
https://arxiv.org/abs/2601.03192
MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic Memory
https://arxiv.org/abs/2601.03192
MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic Memory
www.arxiv.org/abs/2601.03192
RAGの弱点を受動的なところだと見て、RAGに保存する情報に意図や結果、強化学習における行動価値のようなものを追加する。記憶の検索も類似度と価値を半分ずつ混ぜてランキングづけする。
ベンチマークとしてはHLE, BigCodeBench, ALFWorld, Lifelong Agent Benchが使われている
MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic Memory
www.arxiv.org/abs/2601.03192
RAGの弱点を受動的なところだと見て、RAGに保存する情報に意図や結果、強化学習における行動価値のようなものを追加する。記憶の検索も類似度と価値を半分ずつ混ぜてランキングづけする。
ベンチマークとしてはHLE, BigCodeBench, ALFWorld, Lifelong Agent Benchが使われている
MemRL actually looks really interesting: https://arxiv.org/html/2601.03192v1
It basically adds episodic memory to LLMs.
It achieves this by storing memories as sets of vectors, so rather than with Retrieval Augmented Generation which relies on textual queries, it actually […]
MemRL actually looks really interesting: https://arxiv.org/html/2601.03192v1
It basically adds episodic memory to LLMs.
It achieves this by storing memories as sets of vectors, so rather than with Retrieval Augmented Generation which relies on textual queries, it actually […]
A new technique developed by researchers at Shanghai Jiao Tong University and other institutions enables large language model agents to learn new skills without the need for expensive fine-tuning. Researchers propose memrlA…
A new technique developed by researchers at Shanghai Jiao Tong University and other institutions enables large language model agents to learn new skills without the need for expensive fine-tuning. Researchers propose memrlA…